The study of CH₄ emission flux in rice fields is an important part and one of the hot issues in the current carbon cycle research of terrestrial ecosystems. This paper was devoted to identifying diurnal change patterns of CH₄ emission fluxes by using eddy covariance (EC) measurement, and to analyzing the linkages of CH₄ fluxes with different biophysical factors on a diurnal timescale based on the spectral Granger causality analysis method in Jurong Ecological Experimental Station (JREES) in 2016. The results show that: the CH₄ fluxes showed distinguishable diurnal variations with single peaks during 13:00-16:00 h local time. In the non-rice growing season, the CH₄ fluxes were very low, and the diurnal variation characteristics were not obvious. At the diurnal timescale, CH₄ fluxes reached daily maximums at 14:30 h, later than gross primary productivity (GPP) (11:30 h) and latent heat flux (LE) (14:00 h), and earlier than air temperature (TA) (15:30 h) and soil temperature (TS) (16:30 h). Wavelet coherence has high periodic coherence at the 24 h and 12 h timescales and lower periodic coherence at the weekly timescales between TA, TS, and CH₄ fluxes throughout the rice-growing season. GPP regulates CH₄ fluxes after accounting for the effects of LE, TA, and TS on CH₄ fluxes. LE mirrored the diurnal pattern of CH₄ fluxes when the effects of TA and TS on CH₄ fluxes were considered. Therefore, GPP and LE related to photosynthesis and evapotranspiration are more important factors affecting the diurnal variation of CH₄ fluxes in rice fields than air temperature and soil temperature. This study can provide a causal mechanism explanation for constructing a CH₄ emission model from rice fields.